A netmask that contains 1-bits after 0-bits (the 1-bits are not
contiguous on the left side) is invalid [1] [2].
The code before this PR used to parse and accept such
non-left-contiguous netmasks. However, a coming change that will alter
`CNetAddr::ip` to have flexible size would make juggling with such
netmasks more difficult, thus drop support for those.
[1] https://en.wikipedia.org/wiki/Classless_Inter-Domain_Routing#Subnet_masks
[2] https://tools.ietf.org/html/rfc4632#section-5.1
3bd67ba5a4 Test addr response caching (Gleb Naumenko)
cf1569e074 Add addr permission flag enabling non-cached addr sharing (Gleb Naumenko)
acd6135b43 Cache responses to addr requests (Gleb Naumenko)
7cc0e8101f Remove useless 2500 limit on AddrMan queries (Gleb Naumenko)
ded742bc5b Move filtering banned addrs inside GetAddresses() (Gleb Naumenko)
Pull request description:
This is a very simple code change with a big p2p privacy benefit.
It’s currently trivial to scrape any reachable node’s AddrMan (a database of all nodes known to them along with the timestamps).
We do have a limit of one GETADDR per connection, but a spy can disconnect and reconnect even from the same IP, and send GETADDR again and again.
Since we respond with 1,000 random records at most, depending on the AddrMan size it takes probably up to 100 requests for an spy to make sure they scraped (almost) everything.
I even have a script for that. It is totally doable within couple minutes.
Then, with some extra protocol knowledge a spy can infer the direct peers of the victim, and other topological stuff.
I suggest to cache responses to GETADDR on a daily basis, so that an attacker gets at most 1,000 records per day, and can’t track the changes in real time. I will be following up with more improvements to addr relay privacy, but this one alone is a very effective. And simple!
I doubt any of the real software does *reconnect to get new addrs from a given peer*, so we shouldn’t be cutting anyone.
I also believe it doesn’t have any negative implications on the overall topology quality. And the records being “outdated” for at most a day doesn’t break any honest assumptions either.
ACKs for top commit:
jnewbery:
reACK 3bd67ba5a4
promag:
Code review ACK 3bd67ba5a4.
ariard:
Code Review ACK 3bd67ba
Tree-SHA512: dfa5d03205c2424e40a3f8a41af9306227e1ca18beead3b3dda44aa2a082175bb1c6d929dbc7ea8e48e01aed0d50f0d54491caa1147471a2b72a46c3ca06b66f
Before this change, we would analyze the contents of `CNetAddr::ip[16]`
in order to tell which type is an address. Change this by introducing a
new member `CNetAddr::m_net` that explicitly tells the type of the
address.
This is necessary because in BIP155 we will not be able to tell the
address type by just looking at its raw representation (e.g. both TORv3
and I2P are "seemingly random" 32 bytes).
As a side effect of this change we no longer need to store IPv4
addresses encoded as IPv6 addresses - we can store them in proper 4
bytes (will be done in a separate commit). Also the code gets
somewhat simplified - instead of
`memcmp(ip, pchIPv4, sizeof(pchIPv4)) == 0` we can use
`m_net == NET_IPV4`.
Co-authored-by: Carl Dong <contact@carldong.me>
- Move the decision whether to translate an error message to where it is
defined. This simplifies call sites: no more `InitError(Untranslated(...))`.
- Make all functions in `util/error.h` consistently return a
`bilingual_str`. We've decided to use this as error message type so
let's roll with it.
This has no functional changes: no messages are changed, no new
translation messages are defined.
3c1bc40205 Add extra logging of asmap use and bucketing (Gleb Naumenko)
e4658aa8ea Return mapped AS in RPC call getpeerinfo (Gleb Naumenko)
ec45646de9 Integrate ASN bucketing in Addrman and add tests (Gleb Naumenko)
8feb4e4b66 Add asmap utility which queries a mapping (Gleb Naumenko)
Pull request description:
This PR attempts to solve the problem explained in #16599.
A particular attack which encouraged us to work on this issue is explained here [[Erebus Attack against Bitcoin Peer-to-Peer Network](https://erebus-attack.comp.nus.edu.sg/)] (by @muoitranduc)
Instead of relying on /16 prefix to diversify the connections every node creates, we would instead rely on the (ip -> ASN) mapping, if this mapping is provided.
A .map file can be created by every user independently based on a router dump, or provided along with the Bitcoin release. Currently we use the python scripts written by @sipa to create a .map file, which is no larger than 2MB (awesome!).
Here I suggest adding a field to peers.dat which would represent a hash of asmap file used while serializing addrman (or 0 for /16 prefix legacy approach).
In this case, every time the file is updated (or grouping method changed), all buckets will be re-computed.
I believe that alternative selective re-bucketing for only updated ranges would require substantial changes.
TODO:
- ~~more unit tests~~
- ~~find a way to test the code without including >1 MB mapping file in the repo.~~
- find a way to check that mapping file is not corrupted (checksum?)
- comments and separate tests for asmap.cpp
- make python code for .map generation public
- figure out asmap distribution (?)
~Interesting corner case: I’m using std::hash to compute a fingerprint of asmap, and std::hash returns size_t. I guess if a user updates the OS to 64-bit, then the hash of asap will change? Does it even matter?~
ACKs for top commit:
laanwj:
re-ACK 3c1bc40205
jamesob:
ACK 3c1bc40205 ([`jamesob/ackr/16702.3.naumenkogs.p2p_supplying_and_using`](https://github.com/jamesob/bitcoin/tree/ackr/16702.3.naumenkogs.p2p_supplying_and_using))
jonatack:
ACK 3c1bc40205
Tree-SHA512: e2dc6171188d5cdc2ab2c022fa49ed73a14a0acb8ae4c5ffa970172a0365942a249ad3d57e5fb134bc156a3492662c983f74bd21e78d316629dcadf71576800c
Instead of using /16 netgroups to bucket nodes in Addrman for connection
diversification, ASN, which better represents an actor in terms
of network-layer infrastructure, is used.
For testing, asmap.raw is used. It represents a minimal
asmap needed for testing purposes.
The original ORCHID prefix was deprecated as of 2014-03, the new
ORCHIDv2 prefix was allocated by RFC7343 as of 2014-07. We did not
consider the original ORCHID prefix routable, and I don't see any reason
to consider the new one to be either.
-BEGIN VERIFY SCRIPT-
sed --in-place'' --expression='s/NET_TOR/NET_ONION/g' $(git grep -I --files-with-matches 'NET_TOR')
-END VERIFY SCRIPT-
The --in-place'' hack is required for sed on macOS to edit files in-place without passing a backup extension.
In order to prevent mixups, our internal range is never allowed as a resolve
result. This means that no user-provided string will ever be confused with an
internal address.
We currently do two resolves for dns seeds: one for the results, and one to
serve in addrman as the source for those addresses.
There's no requirement that the source hostname resolves to the stored
identifier, only that the mapping is unique. So rather than incurring the
second lookup, combine a private subnet with a hash of the hostname.
The resulting v6 ip is guaranteed not to be publicy routable, and has only a
negligible chance of colliding with a user's internal network (which would be
of no consequence anyway).
Fix two CSubNet constructor problems:
- The use of `/x` where 8 does not divide x was broken, due to a
bit-order issue
- The use of e.g. `1.2.3.4/24` where the netmasked bits in the network
are not 0 was broken. Fix this by explicitly normalizing the netwok
according to the bitmask.
Also add tests for these cases.
Fixes#6179. Thanks to @jonasschnelli for reporting and initial fix.
Make sure that chainparams and logging is properly initialized. Doing
this for every test may be overkill, but this initialization is so
simple that that does not matter.
This should fix the travis issues.